add range of Model vs Data usage ratio per appointment type to Figure 4_Model_vs_Real_average_annual_usage_by_appt_type_[All_Facility_Levels]

[BinglingICL]

In the current plot, the Model vs Data usage ratio of each appointment type has only one data point. When summarising/extracting the simulation results, we set the parameter only_mean = True.

We can set only_Mean = False and add a range of usage ratios for each appointment type in this plot. NB. The range should include lower (based on 0.025-quantile of simulation usage), mean, and upper values (0.975-quantile).

Make a simple branch to fix this.

[BinglingICL]

Initial trying results:

The error bars for most appointment types are not visible, due to the data scale (from 0.1 to 10.0, so that we set log scale for y axis for a clear plot). Once we calibrate all appointment's ratios to [0.8, 1.2] (for example), will improve this draft plot.

[tbhallett]

@BinglingICL - now fixed, right?

[BinglingICL]

Yes, as fixed in PR #1025 : )